A dynamic address computation mechanism for use in database management

Abstract

A dynamic address computation mechanism for use in database management is presented. The address computation technique allows information structured according to a given pre-defined tree to be stored without any address pointers, yet the populated tree is assumed to have an arbitrary number of branches below non-terminating tree nodes, and have arbitrary length information stored at the tree leaves. The address computation is accomplished using a compact (yet complete) address-free description of the populated tree structure. Utilization of this description allows information at any level of the tree to be directly accessed. The address computation provides the basis for an "intermediate" database architecture which embodies many of the advantages of both associative architectures and classical Von Neumann architectures.